Design And Implementation Of Instrument Identification Algorithm Based On Image Processing

With the development of graphic and image technology, the identification accuracy is increasing and the identification speed is faster. Instrument identification system used to modify the traditional instrument has become an effective solution of solving the problem of instrument informationization and has been widely studied and applied. This paper, based on image pre-processing, median filtering, image segmentation, character position and character identification algorithms, mainly aims at the existing image processing module and identifies the instrument automatically.First, this paper changes the color image into the gray image, and adopts the median filtering algorithm to eliminate the noise of the image. Secondly, this paper adopts the localization coordinates cutting algorithm to highlight the details of the instrument. Thirdly, in order to get the binary image, this paper overlay the clipped image on the background image, and computes the optimal threshold based on double accuracy algorithm. Fourthly, we compute the horizontal projection and the vertical projection of the binary image to get the start and the end point of the characters, and cut the characters. At last, Comparing with three kinds of classic character identification algorithms, we use template-matching algorithm to verify the characters and the selected template map. in order to output the instrument character data finally.This paper is developed by Matlab7.0. Through the experiments, we can make sure that the project can be used in electric power energy consumption monitor system etc. This method has a high sensitivity, strong robustness, and full automated operation. In the future, this method has high practical value and broad market prospects.